* fix

* link demo to paddlenlp

Co-authored-by: Nceci3 <ceci3@users.noreply.github.com>

* fix CMake issues of mkldnn quant demo

* remove deleted libs

* update

* add gp_nas

* fix shortcut

* fix when kernel_size in ss

* fix export

* fix dp

* fix export model prune param twice

* add comment

* fix unittest

* fix

* update

* fix docs

* update

* update

* update

* Update mkldnn_quant readme
Co-authored-by: Ncc <52520497+juncaipeng@users.noreply.github.com>

* add trt/x86cpu deploy doc

* add export model and depth

* Support for pruning training graph with auxiliary parameters

* Fix bugs

* Fix unitests.

* fix prune doc

* fix prune demo batchsize

* fix lr (#593)

* fix lr schedule in prune demo (#595)

* fix prune demo batchsize

* fix lr shcedule in prune demo
;
Co-authored-by: Nwanghaoshuang <wanghaoshuang@baidu.com>

* remove softmax from demo/models (#596)

* fix prune demo batchsize

* fix lr shcedule in prune demo
;

* remove softmax from demo/models
Co-authored-by: Nwanghaoshuang <wanghaoshuang@baidu.com>

* fix prune demo log
Co-authored-by: Nwanghaoshuang <wanghaoshuang@baidu.com>

* fix prune doc

* fix prune demo batchsize
Co-authored-by: Nceci3 <ceci3@users.noreply.github.com>

* update

* add ofa api docs

* fix prune demo

* migrage static api

* fix a bug